4.0.0

• Renamed ArchiveMD5Entries to AccessPackFileHashes
• Added support for Blake3 chunk hashes
• Added an overload FindEntry(ReadOnlySpan<char> filePath)
  • Unlike the FindEntry(string) version, this one does not normalize the directory path separator.

3.0.3

• Enabled IsAotCompatible and Nullable

2.0.1

Full Changelog: 2.0.0...2.0.1

• Do not require _dir suffix in chunked vpks
• Removed System.Security.Cryptography.Algorithms
• Added ContinuousIntegrationBuild

2.0.0

Full Changelog: 1.8.0...2.0.0

• .NET 8.0 is now required
• Removed FindEntry overloads
• Reduced allocations in FindEntry
• Updated logo

1.8.0

Full Changelog: 1.7.0...1.8.0

• GetMemoryMappedStreamIfPossible will read files smaller or equal to 4096 bytes into memory instead of memory mapping.
• GetMemoryMappedStreamIfPossible now supports memory mapping files in non-split packages if the package was read from a FileStream or when using Read(file path).
• CRC32 is now calculated with System.IO.Hashing

1.7.0

Full Changelog: 1.6.2...1.7.0

• CRC32 mismatch exceptions will now print expected and actual values.
• Added GetMemoryMappedStreamIfPossible to get package entry file mmap for split vpks

1.6.2

Full Changelog: 1.6.1...1.6.2

• Added ReadEntry overload with user provided byte array
• Allow VPKs where MD5 section is not exactly 48 bytes

1.6.1

Full Changelog: https://github.com/SteamDatabase/ValvePak/compare/1.6.0...1.6.1?w=1

• Removed IsSignatureValid call from VerifyHashes
• Fixed slashes in AddFile
• Added VerifyChunkHashes
• Added VerifyFileChecksums

1.6.0

Full Changelog: https://github.com/SteamDatabase/ValvePak/compare/1.5.0...1.6.0?w=1

• Added Write() method that will save to a vpk file (non-chunked)
• Added AddFile() and RemoveFile() methods
• Optimized hash and signature verification to work on stream rather than reading all data into memory

1.5.0

Full Changelog: https://github.com/SteamDatabase/ValvePak/compare/1.4.0...1.5.0?w=1

• Added OptimizeEntriesForBinarySearch(StringComparison)
  • Call this before Read()
  • This sorts the entries, which allows FindEntry(string fileName) to use BinarySearch to find the entries much faster
  • This also allows FindEntry to search case insensitively if desired
  • This is experimental, and maybe removed in the future
• FindEntry(string directory, string fileName) and FindEntry(string directory, string fileName, string extension) overloads have been deprecated